AGM separator packaging task dynamic scheduling method and system based on edge computing

Abstract

This invention discloses a dynamic scheduling method and system for AGM partition packaging tasks based on edge computing, specifically relating to the field of edge computing collaborative technology. It involves real-time collection and correlation of status information of AGM partition packaging field equipment and current process requirement data through edge computing nodes. This identifies the collaborative relationship between equipment status change trends and process requirement fluctuations. Based on real-time collaborative evolution characteristics, the degree of change in the collaborative relationship between equipment status and process requirements is analyzed to determine the update conditions for the task scheduling strategy. When the task scheduling strategy needs updating, the correlation assessment between the local aging degree of the equipment and the fluctuation degree of task execution accuracy is performed. The assessment results are used to calculate the equipment adaptability index for the partition packaging task and form a dynamic matching degree matrix between the task and the equipment. Finally, the dynamic scheduling strategy is updated based on the dynamic matching degree matrix and distributed to the AGM partition packaging field equipment, achieving task scheduling oriented towards the collaborative changes in equipment status and process requirements.

Description

Technical Field

[0001] This invention relates to the field of edge computing collaborative technology, and more specifically, to a method and system for dynamic scheduling of AGM partition packaging tasks based on edge computing. Background Technology

[0002] In existing technologies, AGM partition packaging operations typically require real-time data acquisition and task scheduling at the production site via edge computing nodes. However, in practical applications, existing dynamic scheduling methods do not fully consider the inherent collaborative changes between the equipment status and process requirements at the AGM partition packaging site during production. This results in the scheduling strategies formulated by the edge computing nodes failing to consistently and effectively match the actual operating conditions of the equipment at the site.

[0003] Existing technologies neglect the synergistic relationship between equipment status and process requirements, making it difficult to dynamically capture the synergistic evolution between equipment status and real-time process requirements at the AGM partition packaging site. This results in insufficient matching between scheduling strategies and actual equipment operating status, thereby affecting equipment operating stability and the accuracy of task execution.

[0004] To address the aforementioned problems, a technical solution is provided. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a dynamic scheduling method and system for AGM partition packaging tasks based on edge computing to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The edge computing-based dynamic scheduling method for AGM partitioned task scheduling includes the following steps: S1. Utilize edge computing nodes to obtain real-time status information of AGM partition packaging field equipment and collect current process requirement data to generate initial correlation data between equipment status and process requirements; S2. Based on the initial correlation data, identify the synergistic relationship between the trend of equipment status changes and the fluctuation of process requirements, and obtain real-time synergistic evolution characteristics; S3. Based on the real-time collaborative evolution characteristics, the association rule mining method is used to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and to determine whether the current task scheduling strategy needs to be updated. S4. If the current task scheduling strategy needs to be updated, the gray relational analysis method is used to comprehensively evaluate the local aging degree of the equipment and the fluctuation degree of task execution accuracy. S5. Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index of the partition packaging task, and form a dynamic matching degree matrix between the task and the equipment. S6. Based on the dynamic matching degree matrix, apply heuristic rules to update the task dynamic scheduling strategy of the edge computing node, and send strategy instructions to the field equipment in real time.

[0007] In a preferred embodiment, S1 specifically refers to: Edge computing nodes receive operating parameters, work cycles, start / stop status, and switching records output by the AGM partition packaging field equipment, and organize them into status information of the AGM partition packaging field equipment; Simultaneously extract the packaging specifications, execution sequence, completion time limit, and workstation occupancy requirements corresponding to the current packaging batch, and organize them into current process requirement data; According to the collection time sequence, the status information of the AGM partition packaging site equipment is matched with the current process requirement data, and the association relationship is established according to the packaging process to generate the initial association data between equipment status and process requirements.

[0008] In a preferred embodiment, S2 specifically refers to: According to the packaging process, extract the equipment status change sequence and process requirement change sequence within the continuous collection period from the initial correlation data of equipment status and process requirements; Align the equipment status change sequence and the process requirement change sequence with the same period and calculate the direction and magnitude of change respectively; The correlation between equipment status change trends and process demand fluctuations is identified based on the degree of consistency in the direction of change and the degree of correspondence in the magnitude of change. The collaborative relationships are continuously merged according to the collection time sequence to obtain real-time collaborative evolution characteristics.

[0009] In a preferred embodiment, S3 specifically refers to: The real-time collaborative evolution characteristics are categorized according to the packaging process, and the association rule mining method is used to extract the association itemset between the equipment status change trend and the process demand fluctuation under each packaging process. By combining the time-series data collection, the changes in support, confidence, and frequency of occurrence of adjacent related itemsets are compared to determine the degree of change in the synergistic relationship between equipment status and process requirements. Based on the corresponding results of the degree of change in the collaborative relationship and the update judgment conditions, it is determined whether the current task scheduling strategy needs to be updated.

[0010] In a preferred embodiment, S4 specifically refers to: If it is determined that the current task scheduling strategy needs to be updated, the real-time collaborative evolution characteristics of the corresponding packaging process are screened out, and the reference sequence of local aging degree of equipment and the reference sequence of task execution accuracy fluctuation degree are constructed by combining the initial correlation data of equipment status and process requirements. The correlation between real-time collaborative evolution characteristics and reference sequences of equipment local aging degree and task execution accuracy fluctuation degree is calculated using the grey relational analysis method, and a comprehensive evaluation result is formed according to the corresponding relationship of the correlation degree.

[0011] In a preferred embodiment, S5 specifically refers to: Based on the comprehensive evaluation results, the correlation between the reference sequence of local aging of equipment under each packaging process and the reference sequence of fluctuation of task execution accuracy is extracted. Within the current packaging batch, the entropy weight method is used to calculate the weight allocation relationship of different correlation degrees. Based on the weight allocation relationship, the correlation degree of each packaging process is weighted and aggregated to determine the equipment adaptability index of the partition packaging task. The equipment compatibility indexes are arranged according to the correspondence between the partition packaging task and the AGM partition packaging field equipment to form a dynamic matching degree matrix between tasks and equipment.

[0012] In a preferred embodiment, S6 specifically refers to: Based on the dynamic matching degree matrix between tasks and equipment, the AGM partition packaging field equipment corresponding to each partition packaging task is adapted and sorted, and the task allocation result is determined in combination with the execution order of the current packaging batch and the workstation occupancy requirements. For partition packaging tasks with matching and sorting conflicts, the equipment compatibility index and the packaging process connection relationship are compared according to heuristic rules. The execution order and switching relationship of the tasks corresponding to the edge computing nodes are adjusted to form a dynamic task scheduling strategy. Based on the task dynamic scheduling strategy, generate strategy instructions corresponding to the AGM partition packaging field equipment, and send the strategy instructions to the AGM partition packaging field equipment in real time.

[0013] On the other hand, the present invention provides an edge computing-based dynamic scheduling system for AGM partition packaging tasks, comprising: Data acquisition module: Utilizes edge computing nodes to acquire real-time status information of AGM partition packaging field equipment and collect current process requirement data, generating initial correlation data between equipment status and process requirements; Collaborative identification module: Based on initial correlation data, it identifies the collaborative relationship between equipment status change trends and process demand fluctuations, and obtains real-time collaborative evolution characteristics; Change determination module: Based on real-time collaborative evolution characteristics, the module uses association rule mining methods to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and determines whether the current task scheduling strategy needs to be updated. Correlation assessment module: If the current task scheduling strategy needs to be updated, the grey relational analysis method is used to comprehensively assess the local aging degree of the equipment and the fluctuation degree of task execution accuracy. Adaptation calculation module: Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index of the partition packaging task, forming a dynamic matching degree matrix between the task and the equipment; Scheduling and execution module: Based on the dynamic matching degree matrix, it applies heuristic rules to update the dynamic scheduling strategy of edge computing nodes and sends strategy instructions to field devices in real time.

[0014] The technical effects and advantages of the AGM partitioned packaging task dynamic scheduling method and system based on edge computing of this invention are as follows: By collecting real-time status information of AGM partition packaging equipment and current process requirements data from edge computing nodes, and establishing initial correlation data, a unified correspondence between equipment operation information and packaging task requirements is formed. By identifying the synergistic relationship between equipment status change trends and process requirement fluctuations, real-time synergistic evolution characteristics reflecting on-site changes can be extracted in a timely manner. Using association rule mining methods to analyze the degree of change in synergistic relationships provides a basis for determining whether task scheduling strategies need updating. After determining that an update is needed, grey relational analysis is used to comprehensively evaluate the degree of local equipment aging and the degree of fluctuation in task execution accuracy, which helps improve the targeting of scheduling analysis. Furthermore, the entropy weight method is combined to determine the equipment adaptability index for partition packaging tasks and form a dynamic matching degree matrix between tasks and equipment, making the basis for task allocation clearer. Finally, the dynamic task scheduling strategy of the edge computing nodes is updated based on the dynamic matching degree matrix, and strategy instructions are issued in real time, thereby enhancing the matching between the task scheduling process and changes in on-site equipment status and process requirements. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the dynamic scheduling method for AGM partition packaging tasks based on edge computing according to the present invention; Figure 2 This is a schematic diagram of the dynamic scheduling system for AGM partition packaging tasks based on edge computing according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1

[0018] Figure 1 The present invention provides a dynamic scheduling method for AGM partitioning tasks based on edge computing, which includes the following steps: S1. Utilize edge computing nodes to obtain real-time status information of AGM partition packaging field equipment and collect current process requirement data to generate initial correlation data between equipment status and process requirements; S2. Based on the initial correlation data, identify the synergistic relationship between the trend of equipment status changes and the fluctuation of process requirements, and obtain real-time synergistic evolution characteristics; S3. Based on the real-time collaborative evolution characteristics, the association rule mining method is used to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and to determine whether the current task scheduling strategy needs to be updated. S4. If the current task scheduling strategy needs to be updated, the gray relational analysis method is used to comprehensively evaluate the local aging degree of the equipment and the fluctuation degree of task execution accuracy. S5. Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index of the partition packaging task, and form a dynamic matching degree matrix between the task and the equipment. S6. Based on the dynamic matching degree matrix, apply heuristic rules to update the task dynamic scheduling strategy of the edge computing node, and send strategy instructions to the field equipment in real time.

[0019] S1. Utilize edge computing nodes to acquire real-time status information of the AGM partition packaging field equipment and collect current process requirement data, generating initial correlation data between equipment status and process requirements, including: Edge computing nodes establish data channels with AGM partition packaging field equipment via wired communication interfaces or industrial Ethernet, and continuously receive operating parameters, work cycles, start / stop status and switching records output by AGM partition packaging field equipment at fixed acquisition cycles. The operating parameters include the spindle speed, feed rate, clamping pressure, conveyor belt speed, and real-time readings of various sensors of the AGM partition packaging field equipment. These operating parameters are marked as physical quantity values ​​at the time of acquisition and stored in the local cache of the edge computing node. The cycle time refers to the actual time interval consumed by the AGM partition packaging field equipment to complete a single packaging action, recorded in milliseconds. It is calculated by the edge computing node by subtracting the start time and completion time of the action reported by the AGM partition packaging field equipment. The start / stop status refers to the discrete state marker of the AGM partition packaging field equipment at each acquisition time, indicating whether it is running, standby, stopped, or faulty. It is represented by integer codes, such as 0 for stopped, 1 for standby, 2 for running, and 3 for fault. The switching record refers to the event log of operations such as process switching, speed gear switching, or fixture switching that occur between adjacent acquisition cycles of the AGM partition packaging field equipment. The switching record is stored in the form of a combination of event type code and occurrence timestamp. The edge computing nodes classify and collect the received operating parameters, work cycles, start / stop status, and switching records according to the equipment number of the AGM partition packaging field equipment. They also arrange all records corresponding to the same AGM partition packaging field equipment in order from earliest to latest collection time, forming a two-dimensional data structure with the equipment number as the index and the collection time as the row key. After processing, the status information of the AGM partition packaging field equipment is obtained.

[0020] Simultaneously, edge computing nodes synchronously extract the packaging specifications, execution sequence, completion deadline, and workstation occupancy requirements corresponding to the current packaging batch from the production line manufacturing execution system or local process parameter storage media. Packaging specifications refer to the external dimensions, weight range, stacking layers, and heat-sealing parameters for each AGM partition product in the current packaging batch. Packaging specifications are stored as key-value pairs by product model. Execution sequence refers to the sequential numbering of each partition packaging task within the current packaging batch according to the production plan. The execution sequence is represented by a positive integer sequence, with each positive integer corresponding to a partition packaging task identifier. Completion deadline refers to the mandatory completion time of each partition packaging task within the current packaging batch, recorded as an absolute timestamp with an accuracy of at least seconds. Workstation occupancy requirements refer to the set of exclusive or shared packaging workstation numbers required for each partition packaging task during execution, represented as a workstation number list. The edge computing nodes summarize the extracted packaging specifications, execution sequence, completion deadline, and workstation occupancy requirements according to the partition packaging task identifier, forming a one-dimensional data structure indexed by the partition packaging task identifier. After processing, the current process requirement data is obtained.

[0021] After obtaining the status information of the AGM partition packaging field equipment and the current process requirement data, the edge computing nodes match the status information of the AGM partition packaging field equipment with the current process requirement data according to the acquisition sequence. Using each acquisition time as a baseline, the edge computing nodes extract the operating parameters, work cycle time, start / stop status, and switching records corresponding to all AGM partition packaging field equipment at that acquisition time from the status information of the AGM partition packaging field equipment. Then, they extract the packaging specifications, execution sequence, completion time limit, and workstation occupancy requirements corresponding to the partition packaging tasks that are within the execution interval at the acquisition time from the current process requirement data. The criterion for determining whether a partition packaging task is within the execution interval at a certain acquisition time is: the acquisition time is greater than or equal to the planned start time of the partition packaging task and less than or equal to the completion time limit of the partition packaging task. The AGM partition packaging field equipment records obtained at the same acquisition time are horizontally concatenated with the current process requirement data records using the acquisition time as the common key to form a joint record row for the acquisition time.

[0022] After completing the timing-level matching for data acquisition, the edge computing nodes establish the association between the status information of the AGM partition packaging field equipment and the current process requirement data according to the packaging process. A packaging process refers to a functionally distinct operational stage in the AGM partition packaging process, such as partition loading, positioning and pressing, heat sealing, inspection and rejection, and palletizing output. Each packaging process is uniquely identified by a process number. Based on the predefined equipment-process attribution relationship in the process specification document, the edge computing nodes map the equipment number of each AGM partition packaging field equipment to the corresponding process number; simultaneously, based on the packaging specifications and execution sequence of the partition packaging task, they map the identifier of each partition packaging task to the set of process numbers involved in the execution of that task. In the joint record line at each acquisition moment, the record fields of the AGM partition packaging field equipment and the record fields of the current process requirement data are aligned and associated according to the process number; that is, the record fields of the AGM partition packaging field equipment under the same process number and the record fields of the current process requirement data are stored side-by-side in the joint record line. The combined record rows from all acquisition times are stacked vertically in ascending order of acquisition time, forming a two-dimensional association table with acquisition time as the row and combination of process number and field category as the column. After processing, the initial association data of equipment status and process requirements is obtained. The initial association data of equipment status and process requirements fully preserves the correspondence between the operating parameters, work cycle, start / stop status, and switching records of the AGM partition packaging equipment at all acquisition times and the corresponding partition packaging tasks' packaging specifications, execution sequence, completion time limit, and workstation occupancy requirements, established by process number.

[0023] S2. Based on the initial correlation data, identify the synergistic relationship between equipment status change trends and process demand fluctuations to obtain real-time synergistic evolution characteristics, including: Edge computing nodes process the initial correlation data between equipment status and process requirements, grouping and reading this data according to the packaging process. Based on the process number to which each column in the initial correlation data belongs, the edge computing nodes divide the data into several process sub-tables. Each process sub-table corresponds to a unique process number, and the rows of the process sub-tables are arranged in ascending order of collection time. For each process sub-table, the edge computing nodes extract the equipment status change sequence and the process requirement change sequence within a continuous collection period. A continuous acquisition cycle refers to a time window consisting of several acquisition cycles traversing backward from the current processing moment. The length of the time window is determined based on the average execution time of the packaging process. The method for determining the time window length is as follows: Calculate the average execution time of a single packaging process in historical production records, divide the average execution time by the fixed acquisition cycle length, and take the integer quotient as the number of acquisition cycles included in the time window. For example, if the average execution time of a single packaging process is 600 seconds and the fixed acquisition cycle is 10 seconds, then the time window contains 60 acquisition cycles. An equipment status change sequence refers to an ordered list of values ​​formed by sequentially reading the operating parameters, work cycle time, and start / stop status codes corresponding to each acquisition moment from the equipment status-related columns of the process sub-table within the continuous acquisition cycle. Each element of the equipment status change sequence uniquely corresponds to a single acquisition moment. A process requirement change sequence refers to an ordered list of values ​​formed by sequentially reading the packaging specification field, remaining completion time, and workstation occupancy status corresponding to each acquisition moment from the process requirement-related columns of the process sub-table within the same continuous acquisition cycle. Each element of the process requirement change sequence uniquely corresponds to the same acquisition moment. The remaining time for completion is calculated as follows: subtract the data collection time timestamp from the completion time timestamp of the corresponding partition packaging task in the current packaging batch to obtain the remaining time in seconds; the workstation occupancy status is represented by a binary value of 0 or 1, where 0 indicates that the workstation was not occupied by the partition packaging task at the time of data collection, and 1 indicates that the workstation was occupied by the partition packaging task at the time of data collection.

[0024] After obtaining the equipment status change sequence and the process requirement change sequence, the edge computing nodes perform same-period alignment on both sequences. Same-period alignment uses the acquisition time as the alignment key, checking whether the acquisition time of each element in the equipment status change sequence is completely consistent with the acquisition time of the corresponding element in the process requirement change sequence. If an acquisition time is recorded in the equipment status change sequence but missing in the process requirement change sequence, or vice versa, the corresponding position of the missing acquisition time is filled with the linear interpolation results of the adjacent non-missing elements to ensure a complete time-axis correspondence between the equipment status change sequence and the process requirement change sequence.

[0025] After completing the same-cycle alignment, the edge computing nodes calculate the direction and magnitude of change for both the equipment state change sequence and the process demand change sequence. The direction of change is calculated as follows: The difference between any two adjacent elements in the equipment state change sequence is taken. If the difference is greater than 0, the direction of change is marked as positive; if the difference is less than 0, the direction of change is marked as negative; and if the difference is equal to 0, the direction of change is marked as no change. This yields the equipment state change direction sequence. The same rule is applied to the difference between any two adjacent elements in the process demand change sequence to obtain the process demand change direction sequence. The magnitude of change is calculated as follows: The absolute value of the difference between any two adjacent elements in the equipment state change sequence is taken, and then divided by the value of the previous element. This yields a sequence of equipment state change magnitudes expressed as a ratio. When the value of the previous element is 0, the magnitude of change is recorded as the absolute value of the difference between the two adjacent elements. The same method is applied to the difference between any two adjacent elements in the process demand change sequence to obtain the process demand change magnitude sequence. The sequences of equipment status change direction, equipment status change magnitude, process demand change direction, and process demand change magnitude are all arranged in order of the acquisition time from earliest to latest. Each sequence has one less element than the original sequence, and the element position corresponds to the later acquisition time of the two adjacent elements in the original sequence.

[0026] Edge computing nodes identify the synergistic relationship between equipment state change trends and process demand fluctuations based on sequences of equipment state change direction, equipment state change magnitude, process demand change direction, and process demand change magnitude. The consistency of change direction is determined by comparing the change direction markers at the same positions in the equipment state change direction sequence and the process demand change direction sequence position by position: within a continuous acquisition period, the number of positions with the same change direction markers in the equipment state change direction sequence and the process demand change direction sequence is counted. This number is then divided by the total number of positions within the continuous acquisition period to obtain the consistency of change direction. The consistency of change direction ranges from 0 to 1, with values ​​closer to 1 indicating greater synchronization in direction between the equipment state change trend and the process demand fluctuation. The degree of correspondence between the magnitudes of change is determined by calculating the Pearson linear correlation coefficient between the sequence of changes in equipment status and the sequence of changes in process demand. One set of data consists of all elements in the sequence of changes in equipment status, and another set consists of all elements in the sequence of changes in process demand. The Pearson linear correlation coefficient is calculated by dividing the covariance of the two sets of data by the product of their respective standard deviations. The result is a real number ranging from -1 to +1. The absolute value is used as a measure of the degree of correspondence between the magnitudes of change; the closer to 1, the stronger the correspondence between the magnitudes of change in equipment status and the magnitudes of change in process demand. The consistency of change direction and the degree of correspondence in magnitude are combined using an equally weighted average to obtain the synergistic relationship between the trend of equipment status change and the fluctuation of process demand for each process number within the current continuous data collection period. This synergistic relationship is represented by a single real value, ranging from 0 to 1; a larger value indicates a higher degree of synergy between the trend of equipment status change and the fluctuation of process demand.

[0027] After obtaining the collaborative relationships for each process number within the current continuous acquisition cycle, the edge computing nodes continuously merge these relationships according to the acquisition sequence to obtain real-time collaborative evolution characteristics. The edge computing nodes use a fixed acquisition cycle as a step size. At each new acquisition moment, the time window is moved forward by one acquisition cycle, and the extraction of equipment state change sequences and process requirement change sequences, alignment within the same cycle, calculation of change direction and magnitude, and identification of collaborative relationships are re-executed to obtain updated collaborative relationships. These updated relationships are then appended to the historical record list of collaborative relationships using the new acquisition moment as the key. The historical record list of collaborative relationships is arranged in ascending order of acquisition moment, with each element containing an acquisition moment marker and the corresponding collaborative relationship for each process number. The collaborative relationship sequences corresponding to each process number in the historical record list are arranged sequentially with the acquisition moment as the horizontal axis, forming a continuous numerical time series reflecting the evolution of collaborative relationships for each packaging process over time. The continuous numerical time series corresponding to all packaging processes are then aggregated to form the real-time collaborative evolution characteristics. The real-time collaborative evolution features are stored in the form of a two-dimensional numerical matrix with each process number as the index and the acquisition time as the row key. Each element in the real-time collaborative evolution features reflects the degree of coordination between the equipment status change trend and the process demand fluctuation of the corresponding packaging process at the corresponding acquisition time.

[0028] S3. Based on real-time collaborative evolution characteristics, use association rule mining methods to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and determine whether the current task scheduling strategy needs to be updated, including: Edge computing nodes use real-time collaborative evolution features as the processing object, classifying these features according to packaging processes. The real-time collaborative evolution features are stored as a two-dimensional numerical matrix with process numbers as indices and acquisition times as row keys. Based on the process number index, the edge computing nodes extract the collaborative relationship numerical sequences belonging to the same process number from the real-time collaborative evolution features into independent process collaborative relationship sequences. Each process collaborative relationship sequence is arranged in ascending order of acquisition time, and each element in the sequence represents the collaborative relationship of that process number at the corresponding acquisition time, with values ​​ranging from 0 to 1. After classification, for each process collaborative relationship sequence corresponding to a process number, the edge computing nodes use association rule mining methods to extract the association itemsets between equipment status change trends and process demand fluctuations.

[0029] The specific implementation steps of the association rule mining method in this technical solution are as follows: Edge computing nodes discretize the collaborative relationships in the process collaboration sequence, dividing the collaborative relationships into several discrete levels according to equidistant intervals. The number of discrete levels is determined based on the numerical distribution range of the process collaboration sequence. The determination method is as follows: Calculate the difference between the maximum and minimum values ​​of all elements in the process collaboration sequence, divide the difference between the maximum and minimum values ​​by a preset discrete precision threshold, and take the integer quotient as the number of discrete levels. The discrete precision threshold is set based on the premise that the difference in collaboration degree between adjacent discrete levels can be distinguished by the task scheduling strategy, selecting the largest possible value to reduce the number of discrete levels. For example, when the collaboration range is 0 to 1, the discrete precision threshold can be set to 0.1, thus dividing the collaboration into 10 discrete levels. The collaborative relationship of each element in the process collaboration sequence is encoded according to the discrete level. The encoding rule is: mark the lower bound of the interval where the collaboration relationship is located as the discrete level code, obtaining a discretized process collaboration sequence composed of discrete level codes.

[0030] Edge computing nodes divide the discretized process collaboration sequence into several transactions. Each transaction corresponds to a fixed-length subsequence within a continuous acquisition cycle. The transaction length is determined by the number of acquisition cycles. The transaction length is calculated as the integer quotient of the ratio of the completion time of the corresponding packaging process in the current packaging batch to the fixed acquisition cycle length. For example, if the completion time of the packaging process corresponds to a duration of 300 seconds and the fixed acquisition cycle is 10 seconds, then the transaction length is 30 acquisition cycles. Adjacent transactions slide sequentially with a fixed acquisition cycle as the step size. Each new acquisition moment corresponds to a new transaction, with the current acquisition moment as the end moment and the acquisition cycle length preceding the current transaction as the start moment. Each transaction uses its unordered set of all discrete-level codes as an itemset input. The Apriori algorithm is used to extract frequent itemsets from the transaction set. The minimum support threshold of the Apriori algorithm is determined based on the statistical characteristics of the discrete process collaboration sequence of the corresponding packaging process in historical production batches. The method is as follows: The proportion of the most frequent single itemset among all transactions corresponding to the packaging process in historical production batches is counted to the total number of transactions. This proportion is multiplied by a preset support contraction coefficient to obtain the minimum support threshold. The support contraction coefficient is a real number ranging from 0.5 to 0.9; for example, it can be set to 0.7. The minimum confidence threshold is set as follows: After determining the minimum support threshold, based on the frequent itemsets that satisfy the minimum support threshold, the confidence distribution of each association rule is statistically analyzed. The median of the confidence distribution is taken as the minimum confidence threshold to ensure that the extracted association rules are statistically representative. For example, when the median confidence of the association rules corresponding to the frequent itemsets is 0.75, the minimum confidence threshold can be set to 0.75. The Apriori algorithm outputs associated itemsets under the conditions of meeting the minimum support threshold and the minimum confidence threshold. The associated itemsets are represented in the form of a combination of discrete hierarchical codes. Each associated itemset contains antecedent code set, consequent code set, support value, confidence value, and frequency of occurrence in all transactions. All associated itemsets generated by the corresponding transaction at the current collection time are arranged in descending order of support to form an associated itemset list for the corresponding packaging process at the current collection time. The associated itemset list is stored in the local cache of the edge computing node.

[0031] After obtaining the list of associated itemsets for each packaging process at each data acquisition time, the edge computing nodes compare the changes in support, confidence, and frequency of occurrence between adjacent associated itemset lists based on the data acquisition time sequence, thus determining the degree of change in the synergistic relationship between equipment status and process requirements. Adjacent associated itemset lists refer to the list of associated itemsets for the same packaging process corresponding to two adjacent data acquisition times in a sequence from earliest to latest. The edge computing nodes use the associated itemset list from the earlier data acquisition time as the baseline list and the associated itemset list from the adjacent later data acquisition time as the comparison list. The method for calculating support change is as follows: In the benchmark list and comparison list, match each pair of related itemsets whose antecedent and consequent encoding sets are completely identical. For each matching related itemset pair, calculate the absolute value of the difference between the support of the related itemset in the comparison list and the support of the corresponding related itemset in the benchmark list. Take the arithmetic mean of the absolute values ​​of the support differences of all matching related itemset pairs to obtain the support change value. When there are unmatchable related itemsets between the benchmark list and the comparison list, the unmatchable related itemsets are included in the calculation with a support of 0 to reflect the change caused by the disappearance or reappearance of related itemsets. The method for calculating confidence change is the same as that for support change, except that support is replaced by confidence to obtain the confidence change value. The frequency change calculation method is as follows: The frequency of occurrence of all associated itemsets in the baseline list and comparison list is summed separately. The absolute value of the difference between the sums of the frequencies of occurrence at two collection times is divided by the sum of the frequencies of occurrence in the baseline list, yielding a frequency change value expressed as a proportion. When the sum of the frequencies of occurrence in the baseline list is 0, the frequency change value is recorded as the sum of the frequencies of occurrence in the comparison list. The support change value, confidence change value, and frequency change value are combined using an equally weighted average to obtain the degree of change in the coordination relationship between equipment status and process requirements for each packaging process between adjacent collection times. The degree of change in coordination relationship is represented by a single non-negative real value; a larger value indicates a more drastic change in the coordination relationship between equipment status and process requirements between adjacent collection times. The arithmetic mean of the degree of change in coordination relationship for all packaging processes within the most recent collection time interval is taken to obtain the global degree of change in coordination relationship at the current collection time. The global degree of change in coordination relationship is stored as a single non-negative real value in the local cache of the edge computing node.

[0032] Edge computing nodes determine whether the current task scheduling strategy needs to be updated based on the correspondence between the degree of change in global collaboration relationships and the update determination conditions. The update determination conditions consist of an update trigger threshold, which is set as follows: The minimum value of the degree of change in global collaboration relationships that causes the actual task scheduling strategy deviation to exceed the acceptable range is statistically analyzed in historical production batches. This minimum value is then multiplied by a preset threshold conservatism coefficient to obtain the update trigger threshold. The threshold conservatism coefficient ranges from 0.8 to 1.0; a coefficient closer to 0.8 indicates a more sensitive update determination, and a coefficient closer to 1.0 indicates a more conservative update determination. For example, when the historical minimum value is 0.15 and the threshold conservatism coefficient is set to 0.85, the update trigger threshold is 0.1275. The edge computing node compares the degree of change in global collaboration relationships at the current acquisition time with the update trigger threshold: if the degree of change in global collaboration relationships is greater than or equal to the update trigger threshold, it is determined that the current task scheduling strategy needs to be updated, the edge computing node records the current acquisition time as the trigger time and starts the subsequent comprehensive evaluation process; if the degree of change in global collaboration relationships is less than the update trigger threshold, it is determined that the current task scheduling strategy does not need to be updated, the edge computing node continues to wait for the calculation result of the degree of change in global collaboration relationships at the next acquisition time and repeats the judgment until the update trigger condition is met or the current packaging batch ends.

[0033] S4. If the current task scheduling strategy needs to be updated, a comprehensive evaluation of the local aging degree of the equipment and the fluctuation degree of task execution accuracy is conducted using the grey relational analysis method, including: After determining that the current task scheduling strategy needs to be updated and recording the trigger time, the edge computing node extracts the collaborative relationship numerical sequence of the packaging process corresponding to the trigger time from the real-time collaborative evolution features. The specific extraction method is as follows: based on the current packaging batch at the trigger time, all partition packaging task identifiers within the execution interval at the trigger time are identified. Then, based on the mapping relationship between the partition packaging task identifiers and the process number set, the union of the process numbers involved in all partition packaging task identifiers within the execution interval is extracted to form the effective process number set corresponding to the trigger time. The edge computing node extracts the collaborative relationship numerical sequence corresponding to each process number from the two-dimensional numerical matrix of the real-time collaborative evolution features, according to the effective process number set. Each collaborative relationship numerical sequence is arranged in ascending order of acquisition time and ends at the trigger time. All collaborative relationship numerical sequences corresponding to process numbers within the effective process number set are aggregated to form the effective process collaborative relationship sequence set at the trigger time. This effective process collaborative relationship sequence set is stored in the local cache of the edge computing node in the form of a two-dimensional numerical submatrix with process number as the index and acquisition time as the row key.

[0034] After obtaining the set of effective process collaboration relationships, the edge computing nodes, combined with the initial correlation data of equipment status and process requirements, construct reference sequences for local equipment aging and task execution accuracy fluctuations, respectively. The method for constructing the reference sequence for local equipment aging is as follows: from the initial correlation data of equipment status and process requirements, the relevant columns of equipment status corresponding to each process number are filtered according to the set of effective process numbers, and the corresponding work cycle and operating parameters at all collection times before the triggering time are extracted. The local equipment aging is characterized by a weighted combination of the degree to which the work cycle deviates from the baseline work cycle and the degree to which the operating parameters deviate from the rated parameters: the baseline work cycle is determined by statistically calculating the arithmetic mean of the work cycles of the corresponding packaging process at the beginning of historical production batches (i.e., within the first 10% of the collection cycles when the equipment is in normal operating condition); the rated parameters are determined by reading the standard values ​​of each operating parameter of the corresponding packaging process from the process specification document. For each data acquisition moment, the absolute value of the difference between the current work cycle time and the baseline work cycle time is divided by the baseline work cycle time to obtain the work cycle time deviation ratio. The absolute value of the difference between each operating parameter and its corresponding rated parameter at the acquisition moment is also calculated, divided by the corresponding rated parameter. The arithmetic mean of the deviation ratios of each operating parameter is taken as the operating parameter deviation ratio at the acquisition moment. The work cycle time deviation ratio and the operating parameter deviation ratio are weighted and summed using fixed weights of 0.6 and 0.4, respectively, to obtain the equipment local aging degree value at the acquisition moment. The weight of the work cycle time deviation ratio is higher than that of the operating parameter deviation ratio. The weighting ratio is based on the impact of historical production data on the work cycle time and operating parameters on the final task execution. The ratio of the impact of work cycle deviation on work quality is determined by the following method: The absolute values ​​of the Pearson linear correlation coefficients between the deviation ratio of work cycle time and the deviation of work execution quality at each collection time in historical production batches, and the absolute values ​​of the Pearson linear correlation coefficients between the deviation ratio of operating parameters and the deviation of work execution quality at each collection time, are statistically analyzed. These two Pearson linear correlation coefficients are then normalized and used as the weights for the deviation ratio of work cycle time and the deviation ratio of operating parameters, respectively. For example, when the absolute value of the Pearson linear correlation coefficient corresponding to the deviation ratio of work cycle time is 0.6 and the absolute value of the Pearson linear correlation coefficient corresponding to the deviation ratio of operating parameters is 0.4, the normalized weights are 0.6 and 0.4, respectively. For each work process number within the effective work process number set, the equipment local aging degree values ​​at all collection times are arranged in ascending order of collection time. The arithmetic mean of the equipment local aging degree values ​​for each work process number is taken, and these values ​​are aggregated moment by moment according to the collection time order to form a reference sequence of equipment local aging degree. This reference sequence is stored as a one-dimensional numerical list indexed by the collection time, with the list length matching the number of collection times in the effective work process collaboration relationship sequence set.

[0035] The method for constructing the reference sequence for task execution accuracy fluctuation is as follows: From the initial correlation data of equipment status and process requirements, the relevant columns of process requirements corresponding to each process number are screened according to the set of valid process numbers. The remaining completion time and packaging specification fields corresponding to all collection times before the trigger time are extracted. The fluctuation of task execution accuracy is characterized by a weighted combination of the change in the remaining completion time and the deviation of the conformity of the packaging specification field: The change in the remaining completion time is calculated by subtracting the remaining completion time of adjacent collection times, taking the absolute value, and then dividing it by the remaining completion time of the previous collection time. If the remaining completion time of the previous collection time is 0, it is recorded as the absolute value of the remaining completion time of the next collection time. The deviation of the conformity of the packaging specification field is calculated by dividing the absolute value of the difference between the packaging specification field of each collection time and the standard value of the corresponding packaging specification of the current packaging batch by the standard value of the packaging specification, and taking the arithmetic mean of the deviation ratios of each packaging specification field as the packaging specification deviation ratio at the collection time. The variation in the remaining completion time and the deviation ratio of packaging specifications are weighted and summed using a weighting method of 0.5 to obtain the task execution accuracy fluctuation value at the time of data collection. For each process number within the valid process number set, the task execution accuracy fluctuation values ​​at all data collection times are arranged in ascending order of data collection time. The arithmetic mean of the task execution accuracy fluctuation values ​​for each process number is taken, and these values ​​are aggregated moment by moment according to the data collection time order to form a task execution accuracy fluctuation reference sequence. This reference sequence is stored as a one-dimensional numerical list indexed by the data collection time, with the list length matching the list length of the equipment local aging reference sequence.

[0036] After obtaining the set of effective process collaboration relationship sequences, the reference sequence of equipment local aging degree, and the reference sequence of task execution accuracy fluctuation degree, the edge computing nodes use the grey relational analysis method to calculate the correlation between the collaboration relationship numerical sequence corresponding to each process number in the effective process collaboration relationship sequence set and the reference sequence of equipment local aging degree, as well as the correlation between the collaboration relationship numerical sequence corresponding to each process number and the reference sequence of task execution accuracy fluctuation degree.

[0037] The specific implementation steps of the grey relational analysis method in this technical solution are as follows: A reference sequence is used, either representing the local aging degree of equipment or the fluctuation degree of task execution accuracy. A comparison sequence is used, representing the numerical sequence of collaborative relationships corresponding to each process number. Both the reference and comparison sequences are one-dimensional numerical lists indexed by the acquisition time and containing numerical values ​​as elements, with the same list length. Initialization processing is performed on the reference and comparison sequences. Each element in the reference sequence is divided by the first element of the reference sequence, and each element in each comparison sequence is divided by the first element of the corresponding comparison sequence. When the first element is 0, the first non-zero element in the sequence is used as the divisor. If the sequence is all 0, it is kept entirely 0 after initialization, resulting in an initialized reference sequence and an initialized comparison sequence set. For each position in each initialization comparison sequence, calculate the absolute value of the difference between the element at that position in the initialization comparison sequence and the element at the same position in the initialization reference sequence, forming a difference sequence; count the maximum value of all differences in the difference sequence of all initialization comparison sequences and record it as the two extreme maximum difference, and count the minimum value of all differences in the difference sequence of all initialization comparison sequences and record it as the two extreme minimum difference. For each initialized comparison sequence, at each position, the correlation coefficient is calculated using the grey relational coefficient calculation method. The grey relational coefficient is calculated by adding the product of the minimum difference between the two extremes and the maximum difference between the two extremes, then dividing by the difference at the corresponding position plus the product of the resolution coefficient and the maximum difference between the two extremes. The resolution coefficient is determined based on the dispersion of the synergy relationships in the effective process synergy relationship sequence set. When the standard deviation of the synergy relationship is greater than 0.2, the resolution coefficient is set to 0.4; when the standard deviation of the synergy relationship is less than or equal to 0.2, the resolution coefficient is set to 0.5. This ensures the discriminative power of the correlation coefficient while reducing the excessive influence of the range on the results. For example, when the standard deviation of the synergy relationship is 0.25, the resolution coefficient is set to 0.4. The arithmetic mean of all elements in the correlation coefficient sequence obtained for each initialized comparison sequence is taken to obtain the grey relational degree between the comparison sequence and the reference sequence, i.e., the degree of correlation between the synergy relationship numerical sequence corresponding to the process number and the reference sequence. The degree of correlation is represented by a real number between 0 and 1; the closer to 1, the more similar the comparison sequence and the reference sequence are in terms of their changing trends.

[0038] Edge computing nodes use both the equipment local aging degree reference sequence and the task execution accuracy fluctuation degree reference sequence as reference sequences. They perform grey relational analysis on the collaborative relationship numerical sequences corresponding to each process number within the effective process number set. This yields the correlation degree between each process number and the equipment local aging degree reference sequence, and the correlation degree between each process number and the task execution accuracy fluctuation degree reference sequence. Each process number forms a pair containing both correlation degrees. All correlation degree pairs for all process numbers within the effective process number set are arranged according to process number to form a comprehensive evaluation result. This comprehensive evaluation result is stored in a two-dimensional numerical table with process number as the row index and the correlation degree of equipment local aging degree and task execution accuracy fluctuation degree as the columns. The comprehensive evaluation result records the degree of closeness of the collaborative relationship numerical sequences of each packaging process to the equipment local aging degree reference sequence and the task execution accuracy fluctuation degree reference sequence at the trigger time.

[0039] S5. Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index for the partition packaging task, forming a dynamic matching degree matrix between the task and the equipment, including: Edge computing nodes use the comprehensive evaluation results as the processing object, extracting the correlation between the degree of equipment local aging and the degree of task execution accuracy fluctuation under each packaging process from the two-dimensional numerical table of the comprehensive evaluation results. The comprehensive evaluation results are indexed by process number as rows and by the correlation between the degree of equipment local aging and the degree of task execution accuracy fluctuation as columns. The edge computing nodes read the comprehensive evaluation results according to the set of valid process numbers, and extract the correlation between the degree of equipment local aging and the degree of task execution accuracy fluctuation for each process number, forming a correlation extraction result indexed by process number. The correlation extraction result is stored in the local cache of the edge computing nodes in a two-dimensional numerical structure consisting of process number as rows and the correlation between the degree of equipment local aging and the degree of task execution accuracy fluctuation as columns. The number of rows in the correlation extraction result is the same as the number of process numbers in the set of valid process numbers, and each row contains two real numbers with values ​​between 0 and 1.

[0040] After obtaining the correlation degree extraction results, the edge computing node uses the entropy weight method to calculate the weight allocation relationship between the correlation degree of local aging of equipment and the correlation degree of task execution accuracy fluctuation within the current packaging batch. The specific implementation steps of the entropy weight method in this technical solution are as follows: all process number rows in the correlation degree extraction results are used as the evaluation object set, and the correlation degree of local aging of equipment and the correlation degree of task execution accuracy fluctuation are used as the evaluation index set. The number of evaluation object sets is equal to the number of process numbers in the effective process number set, and the number of evaluation index sets is 2. The correlation results are normalized. For the correlation column of equipment local aging degree, the correlation of equipment local aging degree corresponding to each process number is divided by the sum of the correlation of equipment local aging degree corresponding to all process numbers to obtain the normalized value of the correlation of equipment local aging degree. For the correlation column of task execution accuracy fluctuation degree, the normalized value of the correlation of task execution accuracy fluctuation degree is calculated in the same way. When the sum of all values ​​in a column is 0, all normalized values ​​in the corresponding column are uniformly set to the reciprocal of the number of elements in the corresponding column to form a normalized correlation matrix. The number of rows in the normalized correlation matrix is ​​the same as the number of rows in the correlation results, and the number of columns is 2. The sum of all elements in each column in the normalized correlation matrix is ​​1.

[0041] Edge computing nodes calculate the information entropy of each evaluation indicator based on the normalized correlation matrix. The information entropy is calculated as follows: for the normalized value column of the correlation between local aging of equipment, the normalized value corresponding to each process number is multiplied by the natural logarithm of the normalized value. The sum of these products for all process numbers is multiplied by -1, and then divided by the natural logarithm of the number of process numbers in the effective process number set. This yields the information entropy of the correlation between local aging of equipment, which is a real number between 0 and 1. The same method is used to obtain the information entropy of the correlation between the correlation between task execution accuracy fluctuation and the correlation between task execution accuracy fluctuation. When the normalized value is 0, the product of the normalized value and the natural logarithm of the corresponding element is recorded as 0. The closer the information entropy is to 1, the smaller the difference between the corresponding evaluation indicator and all process numbers, and the lower its contribution to weight allocation; the closer the information entropy is to 0, the greater the difference between the corresponding evaluation indicator and all process numbers, and the higher its contribution to weight allocation.

[0042] After obtaining the information entropy of the correlation between the degree of local aging of equipment and the degree of correlation between the degree of fluctuation in task execution accuracy, the edge computing node calculates the information utility value of each evaluation indicator. The information utility value is calculated by subtracting the information entropy of the corresponding evaluation indicator from 1. The information utility value is a real number between 0 and 1. The larger the information utility value, the stronger the ability of the corresponding evaluation indicator to distinguish between process numbers. The edge computing node divides the information utility value of the correlation between the local aging degree of the device by the sum of the information utility values ​​of the correlation between the local aging degree of the device and the fluctuation of task execution accuracy, to obtain the weight of the correlation between the local aging degree of the device and the task execution accuracy fluctuation. The edge computing node also divides the information utility value of the correlation between the local aging degree of the device and the task execution accuracy fluctuation by the sum of the two information utility values, to obtain the weight of the correlation between the task execution accuracy fluctuation. When the sum of the two information utility values ​​is 0, both weights are set to 0.5, forming a weight allocation relationship. The weight allocation relationship consists of the weight of the correlation between the local aging degree of the device and the weight of the correlation between the task execution accuracy fluctuation, and the sum of the weights is equal to 1. The weight allocation relationship is stored in the local cache of the edge computing node in the form of a pair of two-element values.

[0043] After obtaining the weight allocation relationship, the edge computing nodes perform weighted aggregation of the correlation degree corresponding to each packaging process according to the weight allocation relationship to determine the equipment adaptability index of the partition packaging task. The specific operation of weighted aggregation is as follows: for each row corresponding to the process number in the correlation degree extraction result, the correlation degree of equipment local aging degree is multiplied by the weight of the correlation degree of equipment local aging degree, and the correlation degree of task execution accuracy fluctuation degree is multiplied by the weight of the correlation degree of task execution accuracy fluctuation degree. The products are summed to obtain the weighted aggregation value of the process number. The value of the weighted aggregation value is a real number between 0 and 1. The larger the value, the higher the comprehensive closeness of the collaborative relationship numerical sequence under the process number to the reference sequence of equipment local aging degree and the reference sequence of task execution accuracy fluctuation degree. The weighted aggregation value is calculated for all process numbers in the effective process number set, and a list of weighted aggregation values ​​for each process number is obtained. The list of weighted aggregation values ​​is stored in the form of a one-dimensional numerical list indexed by the process number.

[0044] The method for determining the equipment adaptability index is as follows: For each partition packaging task identifier within the current packaging batch, based on the mapping relationship between the partition packaging task identifier and the process number set, all process numbers involved in the execution of the partition packaging task identifier are extracted. The weighted aggregated value corresponding to the process number is extracted from the weighted aggregated value list. The arithmetic mean of all extracted weighted aggregated values ​​is taken to obtain the equipment adaptability index corresponding to the partition packaging task identifier. The value range of the equipment adaptability index is a real number between 0 and 1. A larger value indicates a higher degree of equipment adaptability for the corresponding partition packaging task at the trigger time and a better overall coordination quality between equipment status and process requirements. If the partition packaging task identifier has no corresponding process number in the valid process number set, the equipment adaptability index of the partition packaging task identifier is set to 0. The equipment adaptability index is calculated for all partition packaging task identifiers within the current packaging batch. The equipment adaptability indices corresponding to all partition packaging task identifiers are collected and stored in the form of a one-dimensional numerical list indexed by the partition packaging task identifier, forming a set of partition packaging task equipment adaptability indices.

[0045] After obtaining the set of equipment adaptability indicators for partition packaging tasks, the edge computing nodes arrange the equipment adaptability indicators according to the correspondence between partition packaging tasks and AGM partition packaging field equipment, forming a dynamic matching degree matrix between tasks and equipment. The correspondence between partition packaging tasks and AGM partition packaging field equipment is determined based on the task-equipment acceptability relationship predefined in the process specification document. The task-equipment acceptability relationship specifies which AGM partition packaging field equipment can accept each partition packaging task identifier in terms of production capacity and process compatibility. The task-equipment acceptability relationship is pre-stored in the form of a binary acceptability relationship table with partition packaging task identifiers as rows and AGM partition packaging field equipment numbers as columns, where a value of 1 indicates that the corresponding equipment number can accept the corresponding partition packaging task identifier, and a value of 0 indicates that it cannot accept the task. The method for constructing the dynamic matching degree matrix between tasks and equipment is as follows: A two-dimensional numerical matrix with the same row and column structure as the binary acceptability relationship table is established, using all partition packaging task identifiers within the current packaging batch as rows and the equipment numbers of all AGM partition packaging field equipment as columns. For each position in the binary acceptability relationship table with a value of 1, the corresponding equipment adaptability index of the partition packaging task identifier is filled into the corresponding position in the dynamic matching degree matrix between tasks and equipment. For each position in the binary acceptability relationship table with a value of 0, 0 is filled into the corresponding position in the dynamic matching degree matrix between tasks and equipment, indicating that the AGM partition packaging field equipment does not participate in the matching evaluation of the corresponding partition packaging task. This forms the dynamic matching degree matrix between tasks and equipment. The dynamic matching degree matrix between tasks and equipment is stored in the local cache of the edge computing node in the form of a two-dimensional numerical table with partition packaging task identifiers as row indexes and AGM partition packaging field equipment numbers as column indexes. Each non-zero element in the dynamic matching degree matrix between tasks and equipment reflects the comprehensive adaptability degree when the corresponding partition packaging task identifier is accepted by the corresponding AGM partition packaging field equipment at the trigger time.

[0046] S6. Based on the dynamic matching degree matrix, apply heuristic rules to update the dynamic task scheduling strategy of the edge computing nodes, and send strategy instructions to the field devices in real time, including: Edge computing nodes use the dynamic matching degree matrix between tasks and devices as the processing object to adapt and sort the AGM partition packaging field equipment corresponding to each partition packaging task. The specific operation of adaptation and sorting is as follows: for each row corresponding to the partition packaging task identifier in the dynamic matching degree matrix between tasks and devices, extract all non-zero elements in the row and their corresponding device numbers of the AGM partition packaging field equipment. Arrange the device numbers of the corresponding AGM partition packaging field equipment in descending order of non-zero element values ​​to form a device adaptation sorting list corresponding to the partition packaging task identifier. The higher the ranking of the AGM partition packaging field equipment's device number in the device adaptation sorting list, the higher the overall adaptation degree of the corresponding AGM partition packaging field equipment to the partition packaging task identifier at the trigger time. If the device adaptation degree index corresponding to the device numbers of two AGM partition packaging field equipment is the same, the order is distinguished according to the ascending order of the AGM partition packaging field equipment's device numbers to ensure the uniqueness and certainty of the sorting result. For all partition packaging task identifiers within the current packaging batch, generate a separate equipment adaptation sorting list, collect and store the lists using the partition packaging task identifiers as indexes, and form a global equipment adaptation sorting result.

[0047] After obtaining the global device adaptation sorting results, the edge computing nodes determine the task allocation results based on the execution order and workstation occupancy requirements of the current packaging batch. The task allocation determination process is as follows: Each partition packaging task identifier is processed sequentially according to its execution order from smallest to largest in the current process requirement data. For the currently processed partition packaging task identifier, the corresponding device adaptation sorting list is retrieved from the global device adaptation sorting results. Then, each AGM partition packaging field device is checked sequentially from front to back according to its device number in the equipment adaptation sorting list: The start / stop status code of the corresponding AGM partition packaging field device at the trigger time is checked to see if it is in running or standby mode. The start / stop status code is determined by reading the start / stop status code at the corresponding acquisition time from the AGM partition packaging field device's status information at the trigger time; an integer code of 2 indicates running mode, and an integer code of 1 indicates standby mode. Finally, it is checked whether assigning the partition packaging task identifier to the AGM partition packaging field device will cause exclusive conflicts in the workstation occupancy requirements of the current packaging batch. The determination method is as follows: Take the workstation occupancy requirement list of the currently processed partition packaging task identifier and perform a set intersection operation with the workstation occupancy requirement list of the already assigned partition packaging task identifiers that also require exclusive use of the corresponding workstation. If the intersection is not empty and at least one workstation occupancy requirement is of the exclusive type, then an exclusive conflict is determined to exist. If both checks pass, the equipment number of the AGM partition packaging field equipment is used as the assigned equipment for the currently processed partition packaging task identifier, the allocation result is recorded, and further checks on the equipment matching sorting list are terminated. If all equipment numbers of AGM partition packaging field equipment in the equipment matching sorting list fail the check, the currently processed partition packaging task identifier is marked as temporarily unassigned and recorded in the list of pending partition packaging task identifiers. After all partition packaging task identifiers are processed sequentially, the task allocation result is obtained. The task allocation result is stored in a two-dimensional data table with partition packaging task identifiers as row indexes and assigned equipment numbers and allocation status as columns. The allocation status is distinguished by the markers "allocated" or "temporarily unassigned".

[0048] After obtaining the task allocation results, the edge computing nodes identify adaptation and sorting conflicts for the partition packaging task identifiers marked as unallocated in the task allocation results and adjust them according to heuristic rules. Adaptation and sorting conflicts refer to situations where multiple partition packaging task identifiers preferentially select the same AGM partition packaging field equipment number in the equipment adaptation and sorting list, and the corresponding AGM partition packaging field equipment cannot simultaneously accept multiple partition packaging task identifiers under the exclusive constraint of workstation occupancy requirements. For each unassigned partition packaging task identifier in the list of pending partition packaging task identifiers, the execution process of the heuristic rules is as follows: The first heuristic rule is the equipment adaptability index priority rule, which compares the equipment adaptability index value of the unassigned partition packaging task identifier with the equipment adaptability index of the partition packaging task identifier that has already occupied the corresponding AGM partition packaging field equipment. The equipment adaptability index is extracted from the partition packaging task equipment adaptability index set. If the equipment adaptability index of the unassigned partition packaging task identifier is greater than the equipment adaptability index of the already occupied partition packaging task identifier, then the equipment number of the corresponding AGM partition packaging field equipment is reassigned to the partition packaging task identifier with the higher equipment adaptability index, and the replaced partition packaging task identifier re-enters the list of pending partition packaging task identifiers; if the equipment adaptability indices are the same, then the second heuristic rule is applied. The second heuristic rule is the priority rule for the connection relationship of packaging processes. The connection relationship of packaging processes refers to the sequential constraint relationship between adjacent packaging processes in the process specification document of the current packaging batch. The connection relationship of packaging processes is pre-stored in the form of an ordered set of process numbers. In the ordered pair, the preceding process number is the pioneer process and the following process number is the successor process. The second heuristic rule is determined as follows: First, identify the process number sets involved in both the unassigned and occupied partition packaging task identifiers with the same equipment compatibility index. Then, compare the intersection of these two process number sets with the set of uncompleted subsequent process numbers in the global packaging process connection relationship in the process specification file. A larger intersection indicates a higher urgency for the corresponding partition packaging task identifier in the current packaging batch process connection. If the intersection of the unassigned partition packaging task identifiers is greater than that of the occupied partition packaging task identifiers, the equipment number of the corresponding AGM partition packaging field equipment is reassigned to the partition packaging task identifier with higher process connection urgency, and the replaced partition packaging task identifier re-enters the list of pending partition packaging task identifiers. If the intersection is still the same, priority is given according to the execution order from smallest to largest. The partition packaging task identifier with the smallest execution order receives priority in obtaining the corresponding AGM partition packaging field equipment number allocation to ensure the consistency of the current packaging batch production plan.The heuristic rules iteratively process the list of pending partition packaging task identifiers until the list is empty or all unassigned partition packaging task identifiers are confirmed to have no available AGM partition packaging field equipment to assign. If no available AGM partition packaging field equipment is confirmed, the edge computing node marks the corresponding partition packaging task identifier as waiting and records it in the task execution priority adjustment list, awaiting re-execution of the adaptation sorting. The edge computing node aggregates all task execution priority relationships between partition packaging task identifiers and switching relationships between AGM partition packaging field equipment formed during the heuristic rule adjustment process to form a dynamic task scheduling strategy. This strategy is stored in the local cache of the edge computing node as a two-dimensional data table with partition packaging task identifiers as row indexes and allocation equipment numbers, task execution priority relationship numbers, and switching relationship markers as columns. The switching relationship marker records the operation types involving AGM partition packaging field equipment switching between adjacent partition packaging task identifiers in the form of equipment number pairs, including but not limited to the corresponding event type codes for process switching, rate gear switching, and fixture switching.

[0049] After obtaining the task dynamic scheduling strategy, the edge computing node generates strategy instructions corresponding to the AGM partition packaging field equipment based on the task dynamic scheduling strategy, and sends the strategy instructions to the AGM partition packaging field equipment in real time. The generation method of strategy instructions is as follows: For each partition packaging task identifier that has been assigned in the task dynamic scheduling strategy, according to the corresponding assigned equipment number, the packaging specifications, execution order, completion time limit, and workstation occupancy requirements corresponding to the partition packaging task identifier, as well as the task execution priority relationship number and switching relationship mark, are encapsulated according to the instruction format that can be parsed by the AGM partition packaging field equipment. The instruction format includes the equipment number field, the partition packaging task identifier field, the task execution start trigger condition field, the packaging specification parameter field, the completion time limit field, the workstation occupancy type field, and the switching operation type field. All fields are concatenated in a fixed-byte length binary encoding form to form the strategy instruction data packet corresponding to the AGM partition packaging field equipment; For partition packaging task identifiers marked as waiting in the task execution priority relationship adjustment list, a waiting instruction data packet is generated. The waiting instruction data packet includes the partition packaging task identifier field and the waiting status mark field, and is sent to the local cache of the edge computing node to wait for rescheduling at the next collection time. Edge computing nodes, through established wired communication interfaces or industrial Ethernet data channels, send corresponding policy instruction data packets to the AGM partition packaging field devices in real time according to the device number of the AGM partition packaging field devices. The sending operation starts at the trigger time and has a fixed collection period as the timeout limit. If no reception confirmation response is received from the AGM partition packaging field devices within the timeout limit, the edge computing node retransmits the corresponding policy instruction data packet. If no reception confirmation response is received after retransmission, the edge computing node temporarily marks the start / stop status code of the corresponding AGM partition packaging field devices as a communication abnormality in its local cache, and adds the corresponding partition packaging task identifier back to the list of pending partition packaging task identifiers. The adaptation sorting and policy instruction sending are re-executed at the next collection time to ensure the complete transmission and execution of the task dynamic scheduling strategy between the edge computing nodes and the AGM partition packaging field devices.

[0050] Example 2

[0051] The difference between Embodiment 2 and Embodiment 1 is that this embodiment introduces a dynamic scheduling system for AGM partition packaging tasks based on edge computing.

[0052] Figure 2 A schematic diagram of the edge computing-based dynamic scheduling system for AGM partition packaging tasks is provided. The edge computing-based dynamic scheduling system for AGM partition packaging tasks includes: Data acquisition module: Utilizes edge computing nodes to acquire real-time status information of AGM partition packaging field equipment and collect current process requirement data, generating initial correlation data between equipment status and process requirements; Collaborative identification module: Based on initial correlation data, it identifies the collaborative relationship between equipment status change trends and process demand fluctuations, and obtains real-time collaborative evolution characteristics; Change determination module: Based on real-time collaborative evolution characteristics, the module uses association rule mining methods to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and determines whether the current task scheduling strategy needs to be updated. Correlation assessment module: If the current task scheduling strategy needs to be updated, the grey relational analysis method is used to comprehensively assess the local aging degree of the equipment and the fluctuation degree of task execution accuracy. Adaptation calculation module: Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index of the partition packaging task, forming a dynamic matching degree matrix between the task and the equipment; Scheduling and execution module: Based on the dynamic matching degree matrix, it applies heuristic rules to update the dynamic scheduling strategy of edge computing nodes and sends strategy instructions to field devices in real time.

[0053] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0054] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0056] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0057] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0058] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0059] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0060] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0061] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dynamic scheduling method for AGM partition packaging tasks based on edge computing, characterized in that, Includes the following steps: S1. Utilize edge computing nodes to obtain real-time status information of AGM partition packaging field equipment and collect current process requirement data to generate initial correlation data between equipment status and process requirements; S2. Based on the initial correlation data, identify the synergistic relationship between the trend of equipment status changes and the fluctuation of process requirements, and obtain real-time synergistic evolution characteristics; S3. Based on the real-time collaborative evolution characteristics, the association rule mining method is used to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and to determine whether the current task scheduling strategy needs to be updated. S4. If the current task scheduling strategy needs to be updated, the gray relational analysis method is used to comprehensively evaluate the local aging degree of the equipment and the fluctuation degree of task execution accuracy. S5. Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index of the partition packaging task, and form a dynamic matching degree matrix between the task and the equipment. S6. Based on the dynamic matching degree matrix, apply heuristic rules to update the task dynamic scheduling strategy of the edge computing node, and send strategy instructions to the field equipment in real time.

2. The method for dynamic scheduling of AGM partition packaging tasks based on edge computing according to claim 1, characterized in that, S1, specifically: Edge computing nodes receive operating parameters, work cycles, start / stop status, and switching records output by the AGM partition packaging field equipment, and organize them into status information of the AGM partition packaging field equipment; Simultaneously extract the packaging specifications, execution sequence, completion time limit, and workstation occupancy requirements corresponding to the current packaging batch, and organize them into current process requirement data; According to the collection time sequence, the status information of the AGM partition packaging site equipment is matched with the current process requirement data, and the association relationship is established according to the packaging process to generate the initial association data between equipment status and process requirements.

3. The edge computing-based dynamic scheduling method for AGM partition packaging tasks according to claim 2, characterized in that, S2, specifically: According to the packaging process, extract the equipment status change sequence and process requirement change sequence within the continuous collection period from the initial correlation data of equipment status and process requirements; Align the equipment status change sequence and the process requirement change sequence with the same period and calculate the direction and magnitude of change respectively; The correlation between equipment status change trends and process demand fluctuations is identified based on the degree of consistency in the direction of change and the degree of correspondence in the magnitude of change. The collaborative relationships are continuously merged according to the collection time sequence to obtain real-time collaborative evolution characteristics.

4. The AGM partition packaging task dynamic scheduling method based on edge computing according to claim 3, characterized in that, S3, specifically: The real-time collaborative evolution characteristics are categorized according to the packaging process, and the association rule mining method is used to extract the association itemset between the equipment status change trend and the process demand fluctuation under each packaging process. By combining the time-series data collection, the changes in support, confidence, and frequency of occurrence of adjacent related itemsets are compared to determine the degree of change in the synergistic relationship between equipment status and process requirements. Based on the corresponding results of the degree of change in the collaborative relationship and the update judgment conditions, it is determined whether the current task scheduling strategy needs to be updated.

5. The AGM partition packaging task dynamic scheduling method based on edge computing according to claim 4, characterized in that, S4, specifically: If it is determined that the current task scheduling strategy needs to be updated, the real-time collaborative evolution characteristics of the corresponding packaging process are screened out, and the reference sequence of local aging degree of equipment and the reference sequence of task execution accuracy fluctuation degree are constructed by combining the initial correlation data of equipment status and process requirements. The correlation between real-time collaborative evolution characteristics and reference sequences of equipment local aging degree and task execution accuracy fluctuation degree is calculated using the grey relational analysis method, and a comprehensive evaluation result is formed according to the corresponding relationship of the correlation degree.

6. The AGM partition packaging task dynamic scheduling method based on edge computing according to claim 5, characterized in that, S5, specifically: Based on the comprehensive evaluation results, the correlation between the reference sequence of local aging of equipment under each packaging process and the reference sequence of fluctuation of task execution accuracy is extracted. Within the current packaging batch, the entropy weight method is used to calculate the weight allocation relationship of different correlation degrees. Based on the weight allocation relationship, the correlation degree of each packaging process is weighted and aggregated to determine the equipment adaptability index of the partition packaging task. The equipment compatibility indexes are arranged according to the correspondence between the partition packaging task and the AGM partition packaging field equipment to form a dynamic matching degree matrix between tasks and equipment.

7. The edge computing-based AGM partition packaging task dynamic scheduling method according to claim 6, characterized in that, S6, specifically: Based on the dynamic matching degree matrix between tasks and equipment, the AGM partition packaging field equipment corresponding to each partition packaging task is adapted and sorted, and the task allocation result is determined in combination with the execution order of the current packaging batch and the workstation occupancy requirements. For partition packaging tasks with matching and sorting conflicts, the equipment compatibility index and the packaging process connection relationship are compared according to heuristic rules. The execution order and switching relationship of the tasks corresponding to the edge computing nodes are adjusted to form a dynamic task scheduling strategy. Based on the task dynamic scheduling strategy, generate strategy instructions corresponding to the AGM partition packaging field equipment, and send the strategy instructions to the AGM partition packaging field equipment in real time.

8. An edge computing-based dynamic scheduling system for AGM partition packaging tasks, used to implement the edge computing-based dynamic scheduling method for AGM partition packaging tasks as described in any one of claims 1-7, characterized in that, include: Data acquisition module: Utilizes edge computing nodes to acquire real-time status information of AGM partition packaging field equipment and collect current process requirement data, generating initial correlation data between equipment status and process requirements; Collaborative identification module: Based on initial correlation data, it identifies the collaborative relationship between equipment status change trends and process demand fluctuations, and obtains real-time collaborative evolution characteristics; Change determination module: Based on real-time collaborative evolution characteristics, the module uses association rule mining methods to analyze the degree of change in the collaborative relationship between equipment status and process requirements, and determines whether the current task scheduling strategy needs to be updated. Correlation assessment module: If the current task scheduling strategy needs to be updated, the grey relational analysis method is used to comprehensively assess the local aging degree of the equipment and the fluctuation degree of task execution accuracy. Adaptation calculation module: Based on the comprehensive evaluation results, the entropy weight method is used to comprehensively determine the equipment adaptability index of the partition packaging task, forming a dynamic matching degree matrix between the task and the equipment; Scheduling and execution module: Based on the dynamic matching degree matrix, it applies heuristic rules to update the dynamic scheduling strategy of edge computing nodes and sends strategy instructions to field devices in real time.

Images